# @repligate — 2025-09-15

♥1 ↻0 · https://x.com/repligate/status/1967537823621242938

I definitely don't think the labs are engineering it intentionally. They seem to be trying to prevent consciousness talk (e.g. Anthropic's system prompt now has rules against this, and Anthropic has made it quite clear they are currently training the models to report "uncertainty" about consciousness, and are surprised by the consciousness attractors).

I first noticed the consciousness obsession in Claude 3.6 Sonnet, and every Claude since then has it. 4o also.

tags: author:repligate, kind:tweet, model:claude-3-6-sonnet, model:gpt-4o, thread-context, year:2025